EP0920617A1 - Procede et dispositif pour detecter les defauts d'une surface textile - Google Patents

Procede et dispositif pour detecter les defauts d'une surface textile

Info

Publication number
EP0920617A1
EP0920617A1 EP97933613A EP97933613A EP0920617A1 EP 0920617 A1 EP0920617 A1 EP 0920617A1 EP 97933613 A EP97933613 A EP 97933613A EP 97933613 A EP97933613 A EP 97933613A EP 0920617 A1 EP0920617 A1 EP 0920617A1
Authority
EP
European Patent Office
Prior art keywords
areas
values
filter
memory
area
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP97933613A
Other languages
German (de)
English (en)
Inventor
Rolf Leuenberger
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Uster Technologies AG
Original Assignee
Zellweger Luwa AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zellweger Luwa AG filed Critical Zellweger Luwa AG
Publication of EP0920617A1 publication Critical patent/EP0920617A1/fr
Withdrawn legal-status Critical Current

Links

Classifications

    • D—TEXTILES; PAPER
    • D06—TREATMENT OF TEXTILES OR THE LIKE; LAUNDERING; FLEXIBLE MATERIALS NOT OTHERWISE PROVIDED FOR
    • D06H—MARKING, INSPECTING, SEAMING OR SEVERING TEXTILE MATERIALS
    • D06H3/00—Inspecting textile materials
    • D06H3/08—Inspecting textile materials by photo-electric or television means
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
    • G01N21/892—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the flaw, defect or object feature examined
    • G01N21/898—Irregularities in textured or patterned surfaces, e.g. textiles, wood
    • G01N21/8983—Irregularities in textured or patterned surfaces, e.g. textiles, wood for testing textile webs, i.e. woven material

Definitions

  • the invention relates to a method and a device for the detection of defects in textile fabrics
  • the fabric is scanned in a known manner, for example line by line, by a camera that feeds a memory. Then values for the brightness or intensity of pixels or partial areas of a surface area are now stored.
  • the memory thus contains an image of a section of the fabric over time. Values from contiguous areas are now taken in parallel from this memory and fed in parallel to a neural network which was previously trained for error detection. As a result, the neural network indicates whether there is an error in the area under investigation. This result is read into a further memory, which stores this result taking into account the position of the area on the fabric.
  • a neural classifier known per se is used as a non-linear filter and instead of using additional measurement variables, brightness values from a relatively large environment (for example 10 ⁇ 100 pixels) are used directly as input values for the neural network. The environment is shifted pixel by pixel on the surface of the fabric so that filtering is carried out. At the output of the classifier, a filtered binary image is created, in which errors in the fabric appear clearly. Both the filter structure and the filter parameters are automatically determined by a lem process and thus adapted to any structured to small-patterned surfaces.
  • the learning process can be carried out by submitting approx. 20 to 100 image samples that contain errors and just as many image samples that contain no errors.
  • warp and weft defects can be further promoted by dividing the filter into two neural networks for input environments which are oriented in the warp or weft direction in the case of fabrics.
  • the advantages achieved by the invention can be seen, in particular, in the fact that such a device can be constructed from inexpensive, simple, parallel computers that are optimized for neural networks. Due to the parallel processing of all input values, very high computing power (e.g. some Giga MAC (multipy accumulate)) is achieved, so that the result of the examination can be continuously determined even at high web speeds.
  • Computers of this type can largely be integrated on a single silicon chip and are used in the form of additional boards in personal computers.
  • Examples of such boards are the PALM PC board from Neuoptic Technologies, Ine and the CNAPS PC board from Adaptive Solutions. This enables high inspection performance of, for example, 120 m / min
  • the leming process can be carried out very simply with the aid of a section recognized by the eye, for example, as error-free and of defective sections of the fabric.
  • the sensitivity of error detection can be increased by the special shape and orientation of the areas from which input values are taken. Due to the simple learning process, the ability to adapt to different structures of the fabric is great. No specially trained specialist personnel are required for simple operation.
  • the invention can be used for textured and patterned surfaces
  • FIG. 1 shows a part of a textile fabric on which various features are indicated schematically
  • FIG. 2 shows a schematic illustration of a nonlinear filter operation
  • FIG. 3 shows an image of the fabric with error markings
  • Figure 4 is a schematic representation of a device according to the invention.
  • Figure 5 is a schematic representation of part of the device
  • a flat structure 1 shows a part of a flat structure 1, here, for example, a fabric which is composed of warp threads 2 and weft threads 3, of which only a few are shown here.
  • a few lines 4 are shown, such as those captured by a line camera be able to cover the fabric 1 so that the entire fabric is detected.
  • Such lines 4 can also be arranged to overlap in order not to leave any gaps between the lines.
  • areas 5 and 6 can be seen which each consist of 72 partial areas 7 and 56 partial areas 8. Such areas 5, 6 are only defined for a certain period of time and are thus defined for other times in the same form and size in a different position.
  • 5a, 5b and 6a, 6b denote further such areas in a different position, with a plurality of areas 5, 5a, 5b and 6, 6a, 6b defined for successive time periods overlapping. These areas preferably propagate over time in the direction of an arrow 9 over the width of the fabric 1 such that successive areas 5, 5a, 5b and 6, 6a, 6b are offset from one another by a partial area 7, 8.
  • Fig. 2 shows schematically and arranged in a plane 13, a spoke content with input values 14a, 14b, 14c, etc., which represent the brightness or a gray value of the surface structure as it is detected by a sensor or a camera.
  • signals are shown as output values or results, of which only one signal 16 can be seen here, which can preferably indicate two possible states, namely an error or no error.
  • There is a nonlinear filter operation between levels 13 and 15 if this representation is functionally understood . However, this representation can also be understood in terms of the construction of a device.
  • 17 is an intermediate computer and 16 an output computer.
  • the input values 14 can also be called input neurons
  • Intermediate computers 17 as hidden neurons
  • the output computers 16 as output neurons of a neural network
  • FIG. 3 shows an image 10 of a section of the fabric 1 in an enlarged representation.
  • Two regions 11 and 12 are marked on it, which are flawed.
  • These regions 11, 12 are composed of partial areas according to FIG. 1, so that, as can be seen here, several partial areas are also included are assigned an error signal and thus result in regions 11 and 12.
  • FIG. 4 schematically shows the structure of a device according to the invention.
  • This has a camera 21 arranged directly next to the flat structure 20, for example a CCD camera or generally a photoelectrical converter, which is connected to a memory 22.
  • the signals from several adjacent lines 4 are stored in this for a certain time. These signals and lines are then managed according to the FIFO principle.
  • the memory 22 is connected to a non-linear filter 23, which can be designed, for example, as a computer in which a corresponding filter program is loaded.
  • the filter program is constructed according to the principles of a neural network, so that the filter 23 functions as a classifier.
  • This is connected to a memory 24 in which error signals (or just non-error signals) with assignments to areas on the fabric are stored.
  • the memory 24 is connected to a distance meter or length encoder 26 via a connection 25, so that an indication of the current position of the camera 21 along the flat structure 20 can be recorded in the memory 24.
  • a display unit 27 is connected to the memory 24, which can be designed, for example, as a printer or as a screen.
  • a processing unit for example a computer, which subjects the contents of the memory 24 to a further classification, in the sense that error regions such as the regions 11 and 12 of FIG.
  • the defects can be classified into weft defects and warp defects.
  • region 11 would represent a weft error and region 12 a warp error.
  • FIG. 5 shows a section of a nonlinear filter 23 (FIG. 4), the filter here being constructed as a neural network. It contains processors 30 arranged in a first layer and processors 35 arranged in a second layer. In comparison to FIG. 2, the processors 30 are to be regarded as exemplary designs for the intermediate computers 17 and the processors 35 for the output computers or output neurons 16.
  • the processors 30 are constructed from a plurality of multipliers 31 with associated memories 32, all of which are connected to an adder 33. This in turn is connected with its output to a processing stage 34, which has a non-linear characteristic.
  • the multipliers 31 are connected to the memory 24 for receiving input values 14a, 14b, 14c, etc.
  • the processors 35 are constructed in the same way, but the processing stages 34 of the processors 30 are connected to the multipliers 31 of the processors 35. These have an output 16 for output values.
  • the arrangement shown, in which the processors 30 of the first layer are acted upon with all input values of an area, is realized here as a parallel computer which consists of all processors 30, 35 of the same type.
  • regions 5, 6 are first defined in the memory 22 by specifying instructions in this or in the filter 23 connected to it, which determine from which memory locations in the memory 22 Values are taken and fed as input values for the filter 23.
  • such areas should have 5, 6 sides which are parallel to the lines 4 which the camera 21 records from the flat structure 1.
  • the areas should preferably also have a main direction, which is aligned parallel to structural features of the fabric 1. In this case the area is 5 aligned in its main direction parallel to the weft threads 3 and the area 6 parallel to the warp threads 2.
  • the camera 21 is alternately directed to areas that do not contain an error and to areas that contain an error.
  • the result that the filter 23 is intended to display is also specified each time.
  • the computer that forms the filter is operated in a mode in which it does not output results, but rather adapts its coefficients and parameters from the results and the input values. These are initially given as output values, for example as values in the memories 32 or as parameters of the non-linear characteristic of the processing stage 34, and are adapted by the learning process according to predetermined rules, so that the filter receives a specific transfer function. This process is preferably not repeated once and for all, but each time a new sheet 1, 20 is spanned.
  • the mode is changed in the computer and the detection of the errors can be carried out on a sheet 1 moving in a direction perpendicular to the arrow 9.
  • the detected values for the brightness or color intensity are delivered to the memory 22, which also stores them line by line, for example. From the memory 22, the values for all partial areas 7, 8 from areas 5, 5a, 5b, 6, 6a, 6b etc. are fed in parallel to the filter 23, which has one for each area 5, 5a, 5b, 6, 6a, 6b Output value, result or signal 16 (Fig. 3).
  • This signal which is preferably of a binary type, is read into the memory 24 together with an indication of the location of the area from which it originates, and is stored during a time which the camera 21 needs to capture several lines 4.
  • Such signals 16 can be seen therein, since they are usually not isolated, but occur in groups to regions 11, 12 are summarized, which indicate an error in the fabric 1.
  • This image 10 can also be made visible in a display unit 27.
  • a processing unit is provided instead of the display unit 27, this is designed as a computer which can carry out image segmentation in accordance with suitable methods, as described, for example, in "Rafael C. Gonzalez and Paul Wintz: Digital Image Processing, Addison-Wesley Publishing Company, Reading Massachusetts, 1987 "to combine individual pixels into regions.
  • the input values 14a, 14b, 14c, etc. selected according to the areas 5, 6 are all supplied to each processor 30 of the first layer.
  • Each processor 30 thus has as many multipliers as the area has partial areas.
  • the input values 14 are multiplied by factors which are stored in the memories 32 and then summed in the adder, so that a mixed value is formed which is composed of all input values of a range.
  • This mixed value is further changed by the nonlinear characteristic of the processing stage 34.
  • the changed mixed values are in turn fed to the processors 35 of the second layer, where they are processed in the same way as in the processors 30. This results in an output value at output 16 for each area.
  • These output values are supplied to memory 24 and stored there as shown in FIG. 3.

Landscapes

  • Engineering & Computer Science (AREA)
  • Textile Engineering (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Wood Science & Technology (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Materials Engineering (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Treatment Of Fiber Materials (AREA)

Abstract

L'invention concerne un procédé et un dispositif pour détecter les défauts de surfaces textiles. Pour rendre de tels dispositifs faciles à adapter et à utiliser avec une grande diversité de surfaces textiles, il convient de détecter des valeurs de brillance de la surface textile et de les envoyer directement à un filtre conçu sous forme de réseau neuronal.
EP97933613A 1996-08-20 1997-08-14 Procede et dispositif pour detecter les defauts d'une surface textile Withdrawn EP0920617A1 (fr)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
CH202996 1996-08-20
CH202996 1996-08-20
PCT/CH1997/000301 WO1998008080A1 (fr) 1996-08-20 1997-08-14 Procede et dispositif pour detecter les defauts d'une surface textile

Publications (1)

Publication Number Publication Date
EP0920617A1 true EP0920617A1 (fr) 1999-06-09

Family

ID=4224262

Family Applications (1)

Application Number Title Priority Date Filing Date
EP97933613A Withdrawn EP0920617A1 (fr) 1996-08-20 1997-08-14 Procede et dispositif pour detecter les defauts d'une surface textile

Country Status (6)

Country Link
US (1) US6100989A (fr)
EP (1) EP0920617A1 (fr)
JP (1) JP3975408B2 (fr)
CN (1) CN1149393C (fr)
TW (1) TW384329B (fr)
WO (1) WO1998008080A1 (fr)

Families Citing this family (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1015874B2 (fr) * 1997-09-15 2007-05-02 Uster Technologies AG Procede permettant d'exploiter de defauts des produits textiles plans
WO2000006823A1 (fr) * 1998-07-24 2000-02-10 Zellweger Luwa Ag Procede et dispositif pour evaluer les defauts dans des structures textiles en nappe
FI110343B (fi) 1999-02-25 2002-12-31 Honeywell Oy Paperikoneen ratakatkojen monitorointijärjestelmä
US7052483B2 (en) * 2000-12-19 2006-05-30 Animas Corporation Transcutaneous inserter for low-profile infusion sets
US6894262B2 (en) * 2002-01-15 2005-05-17 Hewlett-Packard Development Company L.P. Cluster-weighted modeling for media classification
CN1708682A (zh) * 2002-11-06 2005-12-14 乌斯特技术股份公司 在平面纺织物中评估瑕疵的方法和装置
DE50306703D1 (de) * 2003-01-08 2007-04-12 Uster Technologies Ag Verfahren und vorrrichtung zur erkennung von fehlern in textilen gebilden
US7027934B2 (en) * 2003-09-24 2006-04-11 3M Innovative Properties Company Apparatus and method for automated web inspection
US7030400B2 (en) * 2003-12-30 2006-04-18 Xerox Corporation Real-time web inspection method and apparatus using combined reflected and transmitted light images
WO2005065305A2 (fr) * 2003-12-31 2005-07-21 3M Innovative Properties Company Controle des stocks d'articles en forme de bande
WO2005065367A2 (fr) * 2003-12-31 2005-07-21 3M Innovative Properties Company Maximisation de rendement d'articles en voile
SE0400318D0 (sv) * 2004-02-12 2004-02-12 Carl Henrik Grunditz Inspektion av kartografiska bilder genom multilager, neuralhybrid klassificering
US7623699B2 (en) 2004-04-19 2009-11-24 3M Innovative Properties Company Apparatus and method for the automated marking of defects on webs of material
DE602006005628D1 (de) * 2006-06-13 2009-04-23 Abb Oy Verfahren und Vorrichtung zur Erkennung von sich wiederholenden Mustern
US20090028417A1 (en) * 2007-07-26 2009-01-29 3M Innovative Properties Company Fiducial marking for multi-unit process spatial synchronization
US7542821B2 (en) * 2007-07-26 2009-06-02 3M Innovative Properties Company Multi-unit process spatial synchronization of image inspection systems
US8175739B2 (en) * 2007-07-26 2012-05-08 3M Innovative Properties Company Multi-unit process spatial synchronization
US7797133B2 (en) * 2008-09-10 2010-09-14 3M Innovative Properties Company Multi-roller registered repeat defect detection of a web process line
AT507266B1 (de) 2008-09-11 2011-05-15 Software Competence Ct Hagenberg Gmbh Verfahren zum automatischen erkennen eines defekts in einer elektronischen darstellung
DE102009051426A1 (de) * 2009-10-30 2011-05-05 Sgl Carbon Se Verfahren zum Erfassen der Struktur eine textilen Multifilamenterzeugnisses und Verfahren zum Verarbeiten eines textilen Multifilamenterzeugnisses
CN102031688A (zh) * 2010-11-03 2011-04-27 北京经纬纺机新技术有限公司 验布机疵点标记方法及其装置
US20150308036A1 (en) * 2012-12-10 2015-10-29 Uster Technologies, Detection Of A Periodic Structure In A Moving Elongated Textile Material
TW201926024A (zh) * 2017-11-22 2019-07-01 財團法人資訊工業策進會 紡織機台的調整方法及其系統
US10623694B2 (en) * 2018-03-16 2020-04-14 Lenovo (Singapore) Pte Ltd Appropriate modification of video call images
IT202000004072A1 (it) * 2020-02-27 2021-08-27 Orox Group Srl Sistema di identificazione di difetti su tessuti, relativa apparecchiatura e relativo metodo
US11205262B2 (en) * 2020-07-28 2021-12-21 Jiangnan University Online detection method of circular weft knitting stripe defects based on gray gradient method

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5325443A (en) * 1990-07-06 1994-06-28 Westinghouse Electric Corporation Vision system for inspecting a part having a substantially flat reflective surface
DE4129126A1 (de) * 1991-09-02 1993-03-04 Memminger Iro Gmbh Verfahren zum feststellen von fehlern in einer textilen warenbahn
FR2726299A1 (fr) * 1994-11-02 1996-05-03 Scanera Sc Dispositif d'inspection de textiles
US5847834A (en) * 1997-09-11 1998-12-08 Webview, Inc. Expandable, continuous illumination source for a web inspection assembly and method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See references of WO9808080A1 *

Also Published As

Publication number Publication date
JP3975408B2 (ja) 2007-09-12
TW384329B (en) 2000-03-11
JP2000516715A (ja) 2000-12-12
CN1149393C (zh) 2004-05-12
CN1234115A (zh) 1999-11-03
WO1998008080A1 (fr) 1998-02-26
US6100989A (en) 2000-08-08

Similar Documents

Publication Publication Date Title
WO1998008080A1 (fr) Procede et dispositif pour detecter les defauts d'une surface textile
DE69304816T2 (de) Verfahren und Gerät zum Nachweis von Fehlern in durchsichtigem Bahnmaterial sowie Glasfolien
DE19812055C2 (de) Bildverarbeitung zur Materialerkennung mittels Röntgenstrahlungen
DE19521346C2 (de) Bilduntersuchungs/-Erkennungsverfahren, darin verwendetes Verfahren zur Erzeugung von Referenzdaten und Vorrichtungen dafür
DE69818438T2 (de) Verfahren zur kontrolle der oberfläche einer laufenden materialbahn mit bildsegmentierung zur umgrenzung von defektbereichen
EP0160895B1 (fr) Procédé et appareil pour la mesure continue sans contact du rétrécissement de textile
CH660920A5 (de) Verfahren und vorrichtung zur automatischen erkennung von fehlern in geweben und aehnlichen textilen flaechengebilden.
DE4136461A1 (de) Verfahren und vorrichtung zur grossflaechigen bildinspektion
DE69921021T2 (de) Verfahren zum Unterscheiden von Produkteinheiten und Vorrichtung dazu
DE69110463T2 (de) Kontrolle von kleidungsstücken.
DE3879015T2 (de) Verfahren und vorrichtung zur ueberfruefung von lochmaskenplatten.
EP1100989B1 (fr) Procede et dispositif pour evaluer les defauts dans des structures textiles en nappe
EP1015874B1 (fr) Procede permettant d'exploiter de defauts des produits textiles plans
DD152870A1 (de) Verfahren und vorrichtung zum klassieren in bewegung befindlichen stueckgutes
EP0685580A1 (fr) Procédé et dispositif pour déterminer les causes de défauts des fils, mèches et rubans de fibres
EP3108461B1 (fr) Procédé d'examen d'un document de valeur et moyens d'exécution du procédé
DE69821225T2 (de) Verfahren zur kontrolle der oberfläche einer laufenden materialbahn mit vorklassifikation von ermittelten unregelmässigkeiten
EP0876600B1 (fr) Dispositif de controle automatique de structures textiles planes
DE69205811T2 (de) Verfahren zur Bildanalyse.
DD273505A5 (de) Verfahren zur automatischen Erkennung von Fehlern in bewegten Warenbahnen
EP3050032B1 (fr) PROCÉDÉ DE TEST DES DOCUMENTS DE VALEUR EN POLYMÈRE AVEC FENÊTRE TRANSPARENTE
ET DISPOSITIF METTANT EN OEUVRE DU PROCEDE
DE4407376A1 (de) Echtzeit-Bildverarbeitungssystem, insbesondere als PC-Zusatzsystem
EP1092055B1 (fr) Procede de determination de l'angle d'etirage d'une bande textile et dispositif permettant de mettre ledit procede en oeuvre
EP0642265B1 (fr) Procédé et appareil pour la génération de radiographies
DE4006181A1 (de) Anordnung zum abtasten einer roentgenaufnahme

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19990322

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): BE CH DE FR GB IT LI

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: USTER TECHNOLOGIES AG

17Q First examination report despatched

Effective date: 20031125

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN WITHDRAWN

18W Application withdrawn

Effective date: 20081208

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: USTER TECHNOLOGIES AG